Below is a map of the data we are looking at so far. Pink points are a grid of topographical, abiotic, and plant community surveys. Green dots come from a vegetation survey spanning 40 years, in 1971, ’91, 2001, and 2011. Purple dots are soil moisture and texture locations from data collected by Anna Hermes in 2017.
We are interested in using soil moisture as a predictor of species distributions among alpine plants, but have limited resources for explicitly paired data (i.e. data for a precise point where we have both moisture and community data). To account for this, we are interested in interpolating a surface of estimated moisture (and evenutally other ecological covariates) to act as a predictor for species presence/absence. Initially we were hoping that these surfaces could be interpolated based on spatial autocorrelation - basically averaging between sampled points with kriging. So as an initial exploration, we’ve got a map below showing how moisture is distributed across the 2017 data. (Note that this is across all of our survey dates, so we lose temporal resolution with this map. Plotting all survey dates is feasible, but difficult to pull off in markdown)
Along with moisture data, we also have soil texture data for the 2017 survey. So we make a similar map. Here color denotes the qualitative USDA texture designation of the sample, and the hover text tells you the percentage of each soil type (sand, silt, and clay) a given site had. Our sites only had 4 qualtiative designations: Loam (green), Loamy Sand (orange), Sandy Loam (Blue), and Silty Loam (pink). Pink and orange look very similar on this graph, which is unfortunate, but changing color schemes in plotly is, of course, not as easy as it would seem. Stil. Some nice variability.